Generative AI continues to be one of the most popular topics among our users and our own employees. In a recent webinar Neota explored some of the latest trends and recommendations with generative AI, as well as some of the work that we’ve been doing internally in this area. Since the Neota platform enables integration with any third-party software with a REST API, that, means integration between Neota and many of the LLMs, including through ChatGPT, can be done today.
We’ve been exploring the opportunities these integrations opens up, including running prompts from within a Neota application and pulling in responses – generating anything from application ideas, to creating functional requirement documentation. You can also use it to respond with content review, so if you’re needing a grammar check on your introductory page, or to know whether you’ve created enough test cases to cover all key areas in your QA process.
Within the Neota platform we have a number of methods of manipulating text and creating different outputs such as fillable PDFs and highly tailored Word documents. Combined with generative AI, you can still set up your standard structures and templates within Neota, and then use generative AI to assist with filling in and expanding the content. You can also craft highly tailored prompts and create complex conversations with these AI chatbots like ChatGPT.
One of the challenges with conversing with LLMs is knowing exactly how to ask questions to get the desired response. With a Neota app you can help your users by pre-crafting prompts and only asking users the details that you need from them, with all of the complex work happening behind the scenes.
What we created in our experiment
We used this approach to generate an application – we asked the user only simple questions (name of your app, target audience, and app purpose) and behind the scenes we crafted questions to ChatGPT so that it provided very specific, highly structured responses for us. We then used that information to generate an application in Neota. The questions ChatGPT came up with, and their possible responses were placed on individual forms, with a chosen design applied, and all responses were stored in a database. All with the user answering only 3 questions.
Let’s work together
We believe our work so far is just the beginning and are continuing to explore, as we can see the potential for how it can help our customers hit the ground running when building solutions. So, if you have feedback and ideas in this area – for example, on what sorts of applications, and the functionalities you’d like automatically created for you, get in touch with us as we explore what’s possible!